Corneal endothelial replacement therapies, including cell injection therapy, have recently gained great interest clinically in the treatment of patients with vision loss due to endothelial dysfunction. These therapies have the potential to avoid many side effects, such as scarring, endothelial cell density decrease, inflammation, and glaucoma caused by penetrating keratoplasty. However, these studies raise many questions, such as 1) does the underlying matrix, Descemet's membrane (DM), affect the restoration of endothelial function? 2) Is the DM necessary for cell differentiation? And 3) do the transplanted cells reform new DM? Therefore, we propose to develop a 3-dimensional (3D) cell co- culture model consisting of self-associated human corneal fibroblast (HCF) matrix and human corneal endothelial cells (HCEndoC). This model will allow us to examine human corneal endothelial maturation and DM reformation in a controlled environment that mimics the human cornea. To date, we have made two fundamental observations. First, since high levels of crystallins, such as ALDH1A1, are expressed in corneal epithelium and keratocytes and are associated with corneal development and clarity, we examined expression of ALDH1A1 and found that human corneal endothelium in vivo has higher ALDH1A1 mRNA levels than endothelial cells in culture. In addition, ALDH1A1 levels in endothelial cells were enhanced when cultured on the HCF matrix. The second observation is that thrombspondin 1 (TSP1), which is normally localized in the endothelium and DM, appears to play a role in DM repair as the lack of TSP1 results in persistent edema and wound gaping in a mouse corneal penetrating wound model. We postulate that TSP1 is important in endothelial migration and DM restoration during wound healing. Our overlying hypothesis is that co-culture of HCEndoC on our self-assembled fibroblast matrix will generate a model that mimics human corneal repair seen following endothelial replacement therapy. We propose two specific aims to address the following questions. First, does our stromal matrix provide a good substrate to promote endothelial maturation and DM formation? Second, is TSP1 expression necessary for endothelial migration and Descemet's membrane formation? Immunofluorescence microscopy, qRT-PCR, and protein assays will be the major analysis tools. In addition, laser-capture microdissection, gene knock down, and a specific inhibitor of TSP1 function will be used. Relevance to Public Health-Age- and disease-related loss of human endothelial cells is a major cause of corneal blindness and the most common cause for penetrating or endothelial keratoplasty in the United States. The global shortage of transplant-grade donor corneal tissue makes these therapies difficult. Understanding how HCEndoC can be manipulated to develop a functioning endothelium is vital to the understanding and success of endothelial replacement therapies.